In 60 patients with Major depressive disorder (MDD) receiving 30 sessions of stimulation, responders to active tDCS showed significant increases in resting-state EEG connectivity that tracked with symptom improvement, and a machine learning model predicted response with 81% mean accuracy. The study also found post-treatment connectivity in the optimized 4 mA HD-tDCS group was significantly higher than conventional 2 mA tDCS and sham.

Why It Matters To Your Practice

• EEG-derived network metrics may offer an objective, physiology-linked lens on who is benefiting from neuromodulation—beyond symptom scales alone.

• If response can be predicted early (or pre-treatment) with acceptable performance, clinics could reduce trial-and-error and shorten time to effective care.

• Optimized HD-tDCS (4 mA) showing higher post-treatment connectivity than conventional tDCS suggests protocol choice may materially affect outcomes.

Clinical Implications

• Consider that responders may show measurable increases in network strength, global efficiency, and local efficiency—potential candidates for biomarker-informed monitoring in future workflows.

• When offering tDCS for MDD, the data support that more focal, higher-intensity HD-tDCS protocols could yield stronger neurophysiologic changes than standard 2 mA montages (pending replication and safety/tolerability considerations).

• AI-enabled response prediction (81% mean accuracy with participant-level validation and PCA) is promising, but not yet plug-and-play: performance, generalizability, and calibration across sites/devices remain open questions.

Insights

• At baseline, patients with MDD had reduced EEG connectivity vs. 61 matched healthy controls—supporting a measurable network-level deficit state.

• Connectivity gains were most evident in responders and correlated with depression severity improvement, linking the biomarker to clinical change rather than stimulation exposure alone.

• The study design compared conventional tDCS (n=20), optimized HD-tDCS (n=20), and sham (n=20) over six weeks, enabling both within- and between-protocol contrasts.

The Bottom Line

• For MDD, active tDCS response was associated with increased EEG functional connectivity, and optimized HD-tDCS produced the highest post-treatment connectivity.

• AI models reaching 81% mean accuracy hint at a near-term future where EEG plus machine learning helps triage candidates, personalize protocols, and monitor response—once validated for real-world clinical deployment.